A new (and larger) Chetty study on elite college admissions is released today

CollegeAdmissions_Paper.pdf (opportunityinsights.org)

Study of Elite College Admissions Data Suggests Being Very Rich Is Its Own Qualification - The New York Times (nytimes.com)

As noted by David Leonhardt of The New York Times:

Also from the linked The New York Times article:


I have posted on CC before that I have the sense highly selective colleges have a “diverse” student body comprised of the very rich and the significantly low-income — with virtually no one in between. This study seems to confirm that. I am not sure that leads to having a truly representative group of the best and brightest who can exchange ideas for the betterment of all (if that still is, or ever was, a goal).


This isn’t surprising at all. Isn’t this done by design? The top 1% (according to Fortune magazine) is households who make roughly over 655k/year. That sort of income or disproportionate wealth-to-income (which is rarer still) would be needed to subsidize the half of students who are not paying the COA.

It’s really not a problem for those of us that fit into this category. There are plenty of good schools in price range for families who make 200k, 300k, 400k. None of them are Ivy’s, but neither are the alma maters of most of the people in the working world outside of a couple of investment banks.

The wealth disparity does seem more extreme than 20 years ago, but that is more a reflection of the huge increase in COA over that time frame than any overt admissions strategy change. It takes far more income (or wealth, or both) to pay 85k/year than it did to pay 25k/year.


I’m not sure why the wealthiest should continue to be outsized beneficiaries of elite school admissions. It isn’t a matter of education - students can get a great education at many schools - but of access. As long as the levers of power are concentrated among graduates of schools like Harvard it would behoove us, as a society, to have a more balanced representation of students with different backgrounds and life experiences. Today there is too much sameness which, to me, results in a lack of innovative thought and ideas.


I’ll respond just once- what levers of power are you thinking that Ivy grads have exclusive access to?

I can only think of a few- the Supreme Court, a few select investment banks, what else? I know the VC world has a lot of Ivy business grads, but not so many undergrads.

I also know a few Ivy grads and all of them have colleagues who do not have Ivy degrees. I think that is far more common than the mythical vaunted halls of power.

I’m not passing any sort of judgement on the conclusions of this study. I’m just stating that I am not surprised.


From the same NYTimes article I linked to earlier:


At many of the highly rejective schools, the cost per student of providing an education is greater than the gross COA, meaning the college subsidizes even full pay applicants.

There are many sources, here’s one analysis, from 2019.

How does Bowdoin spend its money? – The Bowdoin Orient.


I saw that, but I don’t think of those numbers as the guarantee that they are often portrayed as. Another way of putting it is that 88% of Fortune 500 CEOs are not Ivy grads, for example.

That’s a tenuous correlation IMO.


So the attendance distribution is going to be different from the admission rate distribution, because the application distribution is not even.

You have to go all the way down to Appendix Table 3 Panel A (page 73) of the full paper to get this information:

But this was the attendance distribution for Ivy Plus students by parental income:

0-20 3.2%
20-40 4.7%
40-60 8.2%
60-70 6.0%
70-80 8.4%
80-90 14.9%
90-95 13.5%
95-96 4.1%
96-97 5.2%
97-98 7.1%
98-99 9.1%
99-99.9 12.8%
Top 0.1% 2.9%

If you norm to the size of subrange, this is consistently increasing with higher parental income. The breakeven point was about 80%, meaning 70-80 was a bit below a proportionate share (8.4% of attendees for a 10% size range), and 80-90 was above a proportionate share (14.9% of attendees in a 10% range).

I think it might also be helpful to summarize this in terms of quintiles, which looks like:

0-20 3.2%
20-40 4.7%
40-60 8.2%
60-80 14.4%
80-100 68.5%

Then the top 20% by deciles:

80-90 14.9%
90-100 53.6%

And finally the top 10% by 5% ranges (the 1% ranges are already basically done):

90-95: 13.5%
95-100: 40.1%

That was in fact never the stated goal. Here, for example, is Harvard’s Mission Statement:

Unless you missed the point, here is an elaboration that actually comes first:

Our mission to educate future leaders is woven throughout the Harvard College experience, inspiring every member of our community to strive toward a more just, fair, and promising world.

OK, so Harvard’s mission is not to bring together a “representative group of the best and brightest who can exchange ideas for the betterment of all.” Harvard’s mission is to “educate future leaders.”

Of course some might believe those should be the same groups. But Harvard has never said it sees things that way, and has never acted like it sees things that way.


They are more likely to hold positions of power in government, be heads of corporations etc. That being said, most Ivy grads go on to have typical careers and lives - no different than their non-Ivy league graduate friends/neighbors.


Exactly. Dale and Krueger data show that affluent students do well career wise not because of the school they went to, but for other reasons, including the connections they have. Said differently, they tend to do well no matter what undergrad school they attended.

Separately, and for all, this is a long study which will take many people some time to read and digest. It’s best that people do this before commenting about its findings, no?


At a deep level, this is a chicken and egg problem.

Harvard says its mission is to educate future leaders. If future leaders are going to be disproportionately drawn from higher income families, even more so the highest income families, then Harvard accomplishing its mission requires it to disproportionately educate students drawn from higher/highest income families.

Still, the data quoted above suggests these schools are at least collectively reinforcing this effect.

But I’d also note we’re just looking at the middle part of a chain that goes from birth through career peak. Children of the highest income parents are starting off with extraordinary resources even before they get to kindergarten, in part to prepare them for school. K-8, they are attending expensive private schools and expensive summer programs and expensive clubs and so on, which again are preparing them for high school. They then attend expensive high schools and more expensive summer programs and expensive clubs and so on that prepare them for college. They then attend expensive colleges and such that prepare them for graduate or professional school or entry-level jobs. They then attend expensive graduate or professional schools or highly-selective entry-level jobs in expensive cities that prepare them for the next step after that. And it just keeps going on and on.

OK, so it is true Harvard and its peers are doing their part in all this. But so are “elite” private high schools, and K-8 schools, and pre-Ks. So are “elite” law schools, business schools, and medical schools. So are “elite” entry-level firms and government offices and NPOs. And so on. All these different institutions are taking the children of the “elite” and preparing them to end up “elite” as well.

And that is how our society works. Not just “elite” colleges.


But what this study is finding is there are different effects in the middle of the curve and the far tail of the curve. In the middle of the curve, there is no notable value-added effect. In the far tail of the curve, they apparently are finding a significant value-added effect.

Of course this means the vast majority of attendees of these colleges are not likely to experience a meaningful value-added effect. But apparently a few will.

1 Like

Providing gift link of NYT article: Study of Elite College Admissions Data Suggests Being Very Rich Is Its Own Qualification - The New York Times


That is absolutely correct.

I sometimes think of this as the sweatshirt theory. Just have the ability to wear the right sweatshirt, and you are set for life.

That sort of crude theory about the value of attending elite colleges has long been debunked. For the vast majority of us, many different sweatshirts will get us where we want to go, if what we want is basically a happy and fulfilling life which includes a good job.

But then people will say, what if I want to be a CEO, or Senator, or President of a university, or whatever?

Well, first of all, most people with such ambitions fail. Even if they attend one of these schools. Second, such people are not necessarily happier or more fulfilled.

But still–yeah, maybe it helps a bit.

Nonetheless, my final point would be this. Even if we as a society do not like the fact a disproportionate number of career peak elites come from elite colleges, I don’t think trying to fix that at the college admissions stage makes sense. I think it makes sense to look at how we are selecting our CEOs, Senators, and so on. Meaning start at the end of the process of filling these “elite” positions, not somewhere buried in the middle.


David Deming, one of the study authors, has a substack newsletter where he will be discussing the study, here

He’s also addressing it on his twitter feed (oops his X feed): https://twitter.com/ProfDavidDeming


In other words, the class comes from those scions of aristocratic inheritance who earned enough merit to have typical excellent stats given their ample opportunities, along with those from the non aristocracy who have earned extraordinary levels of merit to be seen as candidates to join the next generation of the aristocracy?


Isn’t that the way the world is run and has always been run? Why would anyone expect anything different from Harvard and its peer institutions? In a historical context, how much have their methods actually changed?


It is not different from historical practice.

But it is different from the carefully marketed popular perception of being a place mainly for those with top academic merit and achievement, even though that admission bucket is significantly smaller than the entire admission class.



This is really just reconfirming a very old model of these “elite” private colleges. At a high level, at least since they became selective and middle-class/middle-American students started attending in significant numbers, they seem to be structured to bring together young people from families with high social, political, and financial capital, and young people who have high intellectual capital, to their mutual profit.

And there are various studies suggesting this works. Maybe not uniquely as to the Ivy Plus colleges, because this is also going on at other highly selective colleges including LACs and flagship public universities and so on. But basically, it looks like the people who enter these schools high on intellectual capital but low on socioeconomic capital do benefit from mingling with the people who enter with high socioeconomic capital.

However, what is sorta “new” is that these colleges have gotten so selective that most of the high socioeconomic capital students ALSO have high intellectual capital. Largely gone are the days of the white shoe Ivy League students with gentlemen’s Cs. There may be a few exceptions, and definitely higher admit rates, but most of the legacy admits and donor’s list admits and elite private feeder HS admits and so on also have excellent grades and test scores.

And this is consistent with another empirical finding, which is that it is actually less clear that these socioeconomic elites who are also high in intellectual capital are benefiting from this mingling, at least in terms of future power/money prospects. One might hope it is still making them better people, though.

The last nuance I would mention is that I do think for all their self-congratulatory talk of educating future leaders, most of those future leaders are going to be just pretty normal professional class leaders, not oligarchs and such. Like, they may become notable leaders in their academic department, but not necessarily ones most people have heard of. Partners in law firms, but not necessarily ones most people have heard of. Business executives, but not necessarily ones most people have heard of. Important officials in government, but not necessarily ones most people have heard of. Have important positions in creative industries, not not necessarily ones most people have heard of. And so on.

And of course some won’t do any of that. Lots of things can happen to derail people during or after college, and all that can happen to Ivy Plus graduates too.

But still–it isn’t like all that is hurt by this mingling effect either. You might never become a CEO, you might “just” be a law firm partner, but law firm partners sometimes interact with CEOs. And being comfortable in those interactions, comfortable generally with networking with anyone who might be in an important position, is a powerful tool in many professional careers.